Decision Forests, Convolutional Networks and the Models in-Between

TL;DR

This paper explores hybrid models combining decision forests and convolutional neural networks, creating a continuum of models that balance accuracy and efficiency, validated on image classification tasks.

Contribution

It introduces conditional networks, a new family of hybrid models that integrate decision trees with CNN-like structures, enhancing efficiency without sacrificing accuracy.

Findings

Hybrid models achieve comparable accuracy to CNNs.

Significant reduction in computational cost.

Smaller model size with maintained performance.

Abstract

This paper investigates the connections between two state of the art classifiers: decision forests (DFs, including decision jungles) and convolutional neural networks (CNNs). Decision forests are computationally efficient thanks to their conditional computation property (computation is confined to only a small region of the tree, the nodes along a single branch). CNNs achieve state of the art accuracy, thanks to their representation learning capabilities. We present a systematic analysis of how to fuse conditional computation with representation learning and achieve a continuum of hybrid models with different ratios of accuracy vs. efficiency. We call this new family of hybrid models conditional networks. Conditional networks can be thought of as: i) decision trees augmented with data transformation operators, or ii) CNNs, with block-diagonal sparse weight matrices, and explicit data routing functions. Experimental validation is performed on the common task of image classification on both the CIFAR and Imagenet datasets. Compared to state of the art CNNs, our hybrid models yield the same accuracy with a fraction of the compute cost and much smaller number of parameters.